Hepatocellular Carcinoma on Chromosome 13Q
Total Page:16
File Type:pdf, Size:1020Kb
Britsh Joumal of Cancer (1995) 72, 383-385 © 1995 Stockton Press All rights reserved 0007-0920/95 $12.00 Evidence for the presence of two tumour-suppressor genes for hepatocellular carcinoma on chromosome 13q T Kurokil, Y Fujiwara2, S Nakamori3, S Imaoka3, T Kanematsu4 and Y Nakamura' 'Laboratory of Molecular Medicine, Institute of Medical Science, The University of Tokyo, Tokyo; 2Second Department of Surgery, Osaka University Medical School, Osaka; 3Division of Surgery, The Centerfor Adult Disease, Osaka; 4The Second Department of Surgery, Nagasaki University School of Medicine, Nagasaki, Japan. Summary The concept that genetic changes accumulate during development and progression of cancer is widely accepted. Frequent allelic losses at chromosome 13q have been found in hepatocellular carcinomas (HCCs), and a known tumour-suppressor at 13ql4, the retinoblastoma (RB) gene, is thought to be the target of those events. However, no strong evidence has emerged to support a significant role of RB during hepatocarcinogenesis. To investigate the minimal area(s) of loss on chromosome 13q in HCCs, we analysed DNAs isolated from 92 tumours for loss of heterozygosity (LOH) at 13 loci on chromosome 13q, using polymorphic microsatellite markers. In 30 (32.6%) of 92 cases we detected LOH for at least one locus on chromosome 13q and 20 revealed a partial or interstitial deletion of chromosome 13q. Deletion mapping of these 20 tumours indicated two separate commonly deleted regions: one was located in the region including RB and the other was located in the region including the BRCA2 locus. These findings suggest that at least one putative tumour-suppressor gene for HCC other than RB, possibly BRCA2, exists on chromosome 13q. Keywords: hepatocellular carcinoma; loss of heterozygosity; chromosome 13q; retinoblastoma gene; microsatel- lite marker; deletion map The genesis of human cancers is generally a multistep process Materials and methods reflecting cumulative genetic alterations that include activa- tion of oncogenes or inactivation of tumour-suppressor Tumours and DNA preparation genes. losses of heterozygosity We and others have reported HCCs and corresponding non-cancerous liver tissues in hepatocellular carcinomas (HCCs) and implied the Primary were obtained from 92 patients during surgery. Part of each presence of tumour-suppressor genes on chromosomes lp, tissue was fixed with formalin and embedded in paraffin for 5q, lIp, (Wang and Roger, 1988; 4q, 8p, 13q, 16q and 17p histological examination; the remaining moiety was stored at et Ding et al., 1991; Buetow et al., 1989; Tsuda al., 1990; -80C for preparation of DNA. Frozen tissue samples were Fujimori et al., 1991; Murakami et al., 1991; Simon et al., ground to a fine powder in liquid nitrogen, suspended in lysis 1991; Walker et al., 1991; Emi et al., 1992, 1993; Nishida et buffer, treated with proteinase K, and extracted by phenol- al., 1992; Sugimura, 1992; Yeh et al., 1994). However, the chloroform-isoamyl alcohol as described elsewhere (Sato et precise molecular mechanism of development and/or progres- sion of HCCs still remains unclear. al., 1990). LOH on chromosome 13q has been observed frequently in primary cancers of the lung (Weston et al., 1989), breast (Lee LOH analysis et al., 1988), bladder (Cairns et al., 1991), ovary (Sato et al., All 13 markers used in the present study represented 1991) and liver (Wang and Roger, 1988; Murakami et al., polymorphic CA-repeat microsatellite markers: D13S217, 1991; Walker et al., 1991; Nishida et al., 1992). The RB gene, D13S260, D13S267, D13S218, D13S263, D13S126, D13S270, a gene responsible for retinoblastoma, located at 13q14, is D13S172, D13S269, D13S170, D13S265, D13S159 and to most involved in the thought be the likely candidate D13S158 (Gyapay et al., 1994; Wooster et al., 1994; Zhang et carcinogenesis of these cancers. However, one report has al., 1994). Each was amplified by the polymerase chain reac- that the RB gene is probably not the target of the suggested tion (PCR) in 10 volumes of a mixture containing frequent allelic deletions on chromosome 1 3q in ovarian jil 1 x PCR buffer (6.7 mM Tris, 16.6 mM ammonium sulphate, cancers (Kim et al., 1994). Zhang et al. (1994) also reported 6.7 gAM EDTA, 10 mM P-mercaptoethanol), 20 pmol each of of the RB gene in HCCs infrequent somatic mutation unlabelled primer and primer labelled with [y-32P]ATP, 20 ng although nearly half of tumour cells lacked the RB protein. 0.1 U of DNA 250 of Recently, linkage analysis of families carrying hereditary of genomic DNA, Taq polymerase, gM each deoxynucleotide triphosphate and 5 mM magnesium breast cancer localised a second gene responsible for familial chloride. Reactions were performed in 25 cycles under the breast cancer (BRCA2) to 13ql2-13 (Wooster et al., 1994). following conditions: 30 s at 94°C, 30 s at 50-55°C and 30 s presence of a putative tumour- Those results indicated the at 72°C (Gene Amp PCR 9600 System, Perkin Elmer Cetus). suppressor gene other than RB on the long arm of A 5 volume of each solution was denatured and then chromosome 13; this unidentified gene might be involved in ftl electrophoresed in 6% polyacrylamide gel containing 7.7 M carcinogenesis of cancers of several tissues, including HCCs. urea and 32% formamide. Gels were fixed in 5% meth- To putative tumour-suppressor define the location of the anol-5% acetic acid, dried and exposed to X-ray film for gene(s) on 13q, we examined deletion mapping studies of 92 HCCs using highly polymorphic microsatellite markers. Here 8-24h. we present evidence that loss of heterozygosity occurs in two discrete regions of this chromosomal arm. Determination of LOH To evaluate the allelic dosage, the signal intensities of polymorphic alleles were quantified with a Hoefer GS-300 Correspondence: Y Nakamura, Laboratory of Molecular Medicine, scanning densitometer; the peak area corresponding to each Institute of Medical Science, The University of Tokyo, 4-6-1, signal was calculated by electronic integration using a GS Shiroganedai, Minato-ku, Tokyo 108, Japan 370 one-dimensional electrophoresis data system (Hoefer Received 16 January 1995; revised 21 March 1995; accepted 4 April Scientific Instruments). The extent of dosage change of given 1995 allele in HCC was calculated by division of the ratio of Two commonly deleted regions of 13q in HCC T Kuroki et al 384 intensity of the abnormal allele to that of the normal allele in Results tumour DNA by the corresponding ratio measured in non- tumorous DNA. The amount of the tumour and correspon- The linear order of the 13 microsatellite loci analysed for ding normal DNAs for PCR reaction was adjusted by LOH, and the RB locus, had been reported previously as repeated PCR experiments. When >50% reduction in signal follows: centromere D13S217-Dl 3S260-Dl 3S267-D13S218 intensity was detected, it was judged as loss of heterozy- -D13S263-D13S126-D13S270-(RB gene)- D13S172-D13S269 gosity. -D13S170-D13S265-D13S159-D13S158 telomere (Gyapay et al., 1994; Wooster et al., 1994; Zhang et al., 1994). Examples of LOH at several of these loci are shown in Figure 1. All 92 tumours were informative for at least one of the loci examined, and in 30 (32.6%) of them we detected LOH; ten No. 57 No. 239 of these cases showed LOH at all informative loci examined N T N T and 20 cases revealed partial or interstitial deletions of chromosome 13q. Figure 2 summarises the results of LOH analyses in these 20 HCCs. Tumour 171 retained heter- ozygosity at the Dl3S263 locus, but showed LOHs in the region distal to D13S126. Tumour 57 showed LOHs in the region between the D13S263 and D13S269 loci, a region that includes the RB locus, but retained heterozygosity for all other informative loci. Tumours 205 and 239 showed LOHs N T between D13S217 and D13S263, both located proximally to the RB gene locus, but retained heterozygosity at all infor- mative loci more proximal and distal to this region. Of the ttl T tumours with partial losses at 13q, two cases (tumours 205 and 239) showed LOH at loci proximal to the RB gene locus; two cases (tumours 57 and 171) showed LOH at loci around the RB gene locus; and the other cases showed possible LOH in both regions. These results indicate that two separate regions are commonly deleted in HCCs, one between markers D13S263 and D13S172 and the other between D13SD217 and D13S263, where the BRCA2 gene is thought to be located (Wooster et al., 1994). Ni T With respect to pathological data of tumours, LOH on 13q was significantly higher in moderately or poorly differentiated types (26 of 63 informative cases) than in well-differentiated or early carcinomas (2 of 24 cases) (P = 0.002 by Fisher's exact test). Similarly, LOH of 1 3q was higher in T3/T4 tumours (15 of 33 cases) than in TI (2 of 13 cases) or T2 (11 of 41 cases) tumours. Discussion Figure 1 LOH analysis on chromosome 13q for two selected HCCs. Microsatellite loci are identified below each autoradiog- The deletion map constructed in this study implies that two ram of paired DNAs from HCC (T) and corresponding normal separate regions on chromosome 13q contain HCC-assoc- tissue (N). Case 57 shows LOH at D13S270 and retention at iated tumour-suppressor genes. The region between D13S263 D13S263 and D13S269. Case 239 shows LOH at D13S260 and and D13S172 covers a genetic distance of about 15 cM retention at D13S270. (Gyapay et al., 1994) and includes the RB gene (Gyapay et 38 40 42 48 57 60 68 141 155 157 165 171 194 197 199 205 228 239 240 257 (centromere) D13S217 4 4 D13S260 4 4 : BRCA2 D13S267 I D13S218 D13S263 0 D13S126 I 4 D13S270 (RB gene) D13S172 I D13S269 I 4 04 D13S170 I 1 4 04 D13S265 I 4 4 4 44 I 4 D13S159 4 01 D13S158 1 ( 4 (telomere) Figure 2 Schematic representation of partial deletions on chromosome 13q in HCCs.